ON  THE  PEAK  OP  TENERIEPE. 
505 
lowering  in  altitude  only,  all  the  observations  being  made  at  the  same  station,  viz. 
Guajara  at  a height  of  8903  feet. 
A zenith  spectrum  then,  at  that  height  and  with  the  particular  apparatus  employed, 
would  appear  to  begin  between  a and  B ; and,  excepting  the  lines  C and  D,  to  have 
nothing  noticeable  between  B and  E.  A horizontal  spectrum,  on  the  other  hand,  sta- 
tion and  instrument  remaining  the  same,  begins  outside  A,  has  numerous  powerful 
bands  of  hues  between  A and  B,  but  none  between  B and  C ; and  while  C has  not  in- 
creased in  thickness,  B has  more  than  quadrupled  its  size,  maintaining  full  sharpness 
and  definition.  Again,  a certain  excessively  fine  hne,  at  the  distance  of  B to  C beyond 
C,  has  grown  to  many  times  the  thickness  of  the  latter,  and  is  accompanied  by  a broad 
and  marked  band  of  finer  lines.  A more  extensive  increase  still  is  perceived  in  the 
innumerable  lines  between  C and  D,  and  immediately  beyond  D.  Of  these  variations 
from  a zenith  spectrum,  as  observed  phenomena,  there  is  no  doubt;  for  the  series  of 
drawings,  taken  quite  independently,  are  found  on  being  now  brought  together,  to  con- 
firm each  other  in  the  result  of  the  growth  of  these  lines  with  the  zenith  distance.  The 
observations  of  August  9 are  the  most  important  to  this  end : there  were  three  draw'- 
ings  obtained  the  same  evening,  and  the  lines  grew  visibly  under  my  eyes : the  red  end 
apparently  lengthened  out  from  B to  A;  additional  lines  were  seen  every  succeeding 
moment,  and  the  old  ones  became  better  defined,  causing  nebulous  bands  at  85°  zenith 
distance  to  become  groups  of  fine  black  lines  at  91°  zenith  distance.  Looking  at  the 
same  time  to  the  notes  of  colour,  the  red  seems  continually  to  have  grown,  to  the  final 
exclusion,  at  91°  zenith  distance,  of  yellow  as  a pure  colour,  orange  merging  at  once  into 
green. 
Contrasting  now  No.  1 with  Nos.  2 and  3,  we  find,  if  we  may  overlook  the  two  latter 
having  been  observed  by  refiected,  instead  of  direct  rays,  that  a zenith  spectrum  at  the 
sea-level  differs  most  notably  from  one  on  the  mountain,  by  the  greater  extent  of  the 
red  end  visible,  and  by  the  increased  number  of  lines ; and  that,  though  it  shows  some 
approximation  to  a low-sun  mountain  spectrum,  it  has  some  radical  differences  there- 
from. 
Professor  Stokes  indeed  suggests,  that  the  deficiency  of  the  red  end  of  my  high-sun 
mountain  spectra  may  be  due  to  the  more  luminous  rays  extinguishing  the  faint  red ; 
and  that  if  the  former  had  been  absorbed  by  a cobalt-blue  glass,  or  the  extreme  red 
separated  by  refraction  through  a second  prism,  A might  have  been  seen ; just  as  in  fact 
it  was  seen  with  the  low-sun  spectra,  when  the  atmosphere  acted  the  part  of  a suitable 
absorbing  medium  ; and  the  suggestion  is  of  extreme  value  for  future  experiments,  though 
it  is  proper  to  state,  that  on  Teneriffe  the  magnifying  power  employed  was  such,  as  neces- 
sarily to  throw  the  brightest  part  of  the  spectrum  out  of  the  field  of  view  when  the 
extreme  red  was  under  examination. 
Red  End  of  Sky  Spectrum. 
No.  13  appears  to  show  that  a zenith  sky  spectrum  on  the  mountain  with  the  sun  in 
3x2 
